Prediction of unfavorable course of heart failure in patients with concomitant thyroid pathology

Abstract

The aim — to determine the predictors of unfavorable course of heart failure (HF) with account of thyroid status of patients.
 Materials and methods. A comprehensive examination involved 381 patients with HF, developed against the background of coronary heart disease (CHD) after infarction cardiosclerosis. All patients were hospitalized in Cardiology department due to HF decompensation. From them, 57.22 % had comorbid thyroid pathology (TP): 92 (24.15 %) non­toxic goiter (NG) and 126 (33.07 %) autoimmune thyroiditis (AIT). 142 patients (37.3 %) were diagnosed with low triiodothyronine syndrome (LT3S), and 26 (6.82 %) were diagnosed with subclinical hypothyroidism. The results of the study were processed using the statistical package of IBM®SPSS® Statistics, 20.0 and MedCalc, 16.4 (free version). To identify adverse factors for the course of HF in patients with concomitant TP in comparison with patients without this comorbidity, a regression analysis of Cox (proportional risk model (Cox proportional hazards model)) was used. Kaplan­Meyer graphs were constructed.
 Results. The prognostic factor was calculated based on the data of Cox’s regression analysis. With account of β-coefficients of the variables included in the model, the Cox regression equation was developed, which determines the probability of rehospitalization (RH) of HF patients within 24 months. During the ROC analysis it was found that the RH risk in patients with HF due to decompensation of the disease increases when reaching the optimal distribution point for the value of HR > 0.104 (sensitivity — 51.61 %, specificity — 96.47 %, p < 0.0001). The RH risk chance ratio for 2 years was 21.60 (6.34 — 73.63), χ2 = 31.543; p = 0.0001). Repeated ROC analysis revealed that in patients with HF in combination with TP, the risk of RH due to decompensation of the disease increases when reaching the optimal distribution point for the value of HR > 0.043 (sensitivity — 80.77 %, specificity — 74.00 %, p < 0.0001). GHG risk chance ratio for 2 years = 11.95 (3.74 — 38.21), χ2 = 18.35; p = 0.0001). The use of ROC­analysis in the group of patients with CH without TP did not allow to establish a probable prognostic value of the HR index. The model of predicting the adverse course of HF in patients with TP was tested on a separate group of 66 patients with HF against the background of coronary heart disease, which was not included in the development of Cox prognostic regression model. These groups did not differ in age, sex, TP and RH frequency. The sensitivity of the specified value of the coefficient was 86.36 %, specificity = 78.57 %. The prognostic value of the positive result was 86.4 %.
 Conclusions. The cumulative rehospitalization risk in patients with HF against the background of coronary heart disease with concomitant TP is higher compared to patients without this comorbid pathology. The risk is highest in patients with concomitant emergency situations. The high risk of unfavorable HF course in patients with comorbid TP was stipulated by presence of the low triiodothyronine syndrome. According to Cox regression analysis, the following predictors that cause an unfavorable HF course in patients with concomitant TP were defined: levels of low­density lipoprotein cholesterol, interleukin­4 and free triiodothyronine. The risk of unfavorable HF course increases with the value of the hazard ratio: for the regression model > 0.043 (OR = 11.960).